Coronavirus

[aaid]

11 hours ago, Grim Jim said:

Too slow to edit...

Yeah, I did see thplinth's post earlier today, thank you both.   So to me, most of us are going to get it, but we're hoping to slow the time-to-spread so NHS can cope better.    Still to read the Kelly blog, but in the meantime I do like the boosing supply idea...

 

Update:   OK James Kelly doesn't quite buy into this and believes it can be contained.   I'm not convinced (expert that I am not) on this occasion.   We're saying then that every time this happens we can snuff it out with actions drastic enough?

I pay a lot of attention to what James Kelly says on opinion polls and electoral arithmetic.   I'm not sure what his qualifications as an epidemiologist are though.

[mrniaboc]

Some quick statistical sampling:

Number of players + coaches of London EPL teams ~150

Of those there are 2 with confirmed covid 19

That's 1.3%

London's population = 9 million

Estimate of currently infected number in London therefore ~ 120,000

UK science chief reckons there could be up to 10,000 cases UK-wide. I reckon he could be wrong by at least an order of magnitude. 

[Squirrelhumper]

15 hours ago, macy37 said:

Because I’m not in my 50’s or have old parents at risk of dying? Right you are then. 😂

Nah, just your general attention seeking.

[Squirrelhumper]

13 minutes ago, mrniaboc said:

Some quick statistical sampling:

Number of players + coaches of London EPL teams ~150

Of those there are 2 with confirmed covid 19

That's 1.3%

London's population = 9 million

Estimate of currently infected number in London therefore ~ 120,000

UK science chief reckons there could be up to 10,000 cases UK-wide. I reckon he could be wrong by at least an order of magnitude. 

No offence but i'll go with the UK Science chief rather than mrniaboc off TAMB.

[dohadeer]

25 minutes ago, mrniaboc said:

Some quick statistical sampling:

Number of players + coaches of London EPL teams ~150

Of those there are 2 with confirmed covid 19

That's 1.3%

London's population = 9 million

Estimate of currently infected number in London therefore ~ 120,000

UK science chief reckons there could be up to 10,000 cases UK-wide. I reckon he could be wrong by at least an order of magnitude. 

Yeah, that’s a major misunderstanding of statistics you’re showing there.

What are you playing at?

 

Edited March 13, 2020 by dohadeer

[dohadeer]

32 minutes ago, mrniaboc said:

Some quick statistical sampling:

Number of players + coaches of London EPL teams ~150

Of those there are 2 with confirmed covid 19

That's 1.3%

London's population = 9 million

Estimate of currently infected number in London therefore ~ 120,000

UK science chief reckons there could be up to 10,000 cases UK-wide. I reckon he could be wrong by at least an order of magnitude. 

Number of actors who won an Oscar for playing Forrest Gump - one.

Of those, one has the Coronavirus.

Estimate of current percentage of world population with Coronavirus - 100%

 

Doesn’t work like that I’m afraid. Leave the statistics to the experts.

Edited March 13, 2020 by dohadeer

[ParisInAKilt]

33 minutes ago, mrniaboc said:

Some quick statistical sampling:

Number of players + coaches of London EPL teams ~150

Of those there are 2 with confirmed covid 19

That's 1.3%

London's population = 9 million

Estimate of currently infected number in London therefore ~ 120,000

UK science chief reckons there could be up to 10,000 cases UK-wide. I reckon he could be wrong by at least an order of magnitude. 

 

[ParisInAKilt]

Wonder if any Tambers will get it? 

Expect the hospital I’m working at just now to start testing staff and patients next week. 

[mrniaboc]

24 minutes ago, ParisInAKilt said:

 

Lol

[mrniaboc]

45 minutes ago, Squirrelhumper said:

No offence but i'll go with the UK Science chief rather than mrniaboc off TAMB.

Fair enough. I will too. But just pointing out what a quick sample shows. 

Ps: I do have another life off the TAMB

[TDYER63]

15 minutes ago, ParisInAKilt said:

Wonder if any Tambers will get it? 

Expect the hospital I’m working at just now to start testing staff and patients next week. 

I dont usually pay attention to any small cold/ flu symptoms i get as they mostly come to nothing. Yesterday I was in the kitchen and  suddenly went very warm and had a wee cough and thought ‘aw naw, I’ve got it’. 
The NHS will be inundated with calls from folk falsely diagnosing themselves with it. 

[phart]

The majority of folk will get it i imagine. It's just trying to make sure not everyone gets it at once to over stretch the health service.

[mrniaboc]

35 minutes ago, dohadeer said:

Yeah, that’s a major misunderstanding of statistics you’re showing there.

What are you playing at?

 

Is it? Please do explain my major mis-step. I took a sample of the population who are regularly monitored for their health and also at a level of fame that means their ill health is reported, and showed that 1.3% of them is the fraction with coronavirus. 

I guess you could be arguing that professional footballers are more likely to contract the disease. But I can't see why at the moment?

[ParisInAKilt]

Just now, TDYER63 said:

I dont usually pay attention to any small cold/ flu symptoms i get as they mostly come to nothing. Yesterday I was in the kitchen and  suddenly went very warm and had a wee cough and thought ‘aw naw, I’ve got it’. 
The NHS will be inundated with calls from folk falsely diagnosing themselves with it. 

Wonder if they are in the process of creating extra NHS 24 jobs, temporarily to deal with the virus. I know they are doing something similar in Australia. 

[mrniaboc]

31 minutes ago, dohadeer said:

Number of actors who won an Oscar for playing Forrest Gump - one.

Of those, one has the Coronavirus.

Estimate of current percentage of world population with Coronavirus - 100%

 

Doesn’t work like that I’m afraid. Leave the statistics to the experts.

Hahahaha now THAT'S a major misunderstanding of statistics. 

Small numbers for sure, but you'd be better says Hanks is 1 from X amount of Oscar winners, and basing your extrapolation on that percentage.

What was better about my approach is that it's a geographically isolated sample.

[phart]

7 minutes ago, mrniaboc said:

Is it? Please do explain my major mis-step. I took a sample of the population who are regularly monitored for their health and also at a level of fame that means their ill health is reported, and showed that 1.3% of them is the fraction with coronavirus. 

I guess you could be arguing that professional footballers are more likely to contract the disease. But I can't see why at the moment?

You're taking a non-representative sample then grafting it to a full population without making any adjustments for sample size etc.

Not that it matters nor can i be bothered arguing it. You either know how to sample properly or don't.

[TDYER63]

6 minutes ago, ParisInAKilt said:

Wonder if they are in the process of creating extra NHS 24 jobs, temporarily to deal with the virus. I know they are doing something similar in Australia. 

They must be. If the UK can find the money for projects like HS2 then they must be directing cash to this .

Surely .....

[Orraloon]

16 hours ago, Toepoke said:

1 in 15 death rate 

As a comparison regular flu is around 1 in 1000...

 

If you read the link that phart posted yesterday you will see how the number of "true cases" ( ie if you count the day they got infected rather than the day they tested positive) could be as high as 400,000. That could bring down the death rate. But on the other hand a % of those 400,000 will die in the future but haven't been counted yet. We won't get an accurate fatality rate from the number of dead and the number of confirmed cases. Not yet anyway. we need to take account of the lag between the time folk get infected and the time they either die or recover. It does look like it will end up much higher than normal flu though. 

It takes a wee while to read (about 30 minutes) but it explains it very well. 

If I am reading it properly, what it doesn't attempt to take into account is the number of people who get infected but don't get tested. They won't be able to estimate that number with much accuracy until we are well past the exponential phase and things start to settle down.

The eventual fatality rates could be quite different in different regions though, depending on how different countries react.

Go on, read it. It's well worth it. 

 

21 hours ago, phart said:

good involved read

https://medium.com/@tomaspueyo/coronavirus-act-today-or-people-will-die-f4d3d9cd99ca

 

[mrniaboc]

5 minutes ago, phart said:

You're taking a non-representative sample then grafting it to a full population without making any adjustments for sample size etc.

Not that it matters nor can i be bothered arguing it. You either know how to sample properly or don't.

Definitely can be considered a non-representitive sample for sure. But at the moment it's the only sample I have my hands on where I think all cases are being reported, and reported early.

Also defs small number stats that I'm grafting to a large population, so the numbers have large error margins.

But it was a fun thought experiment to do. Say 50% of footballers on london had covid. What estimate would you then make for the general London population? 

Thanks for giving better feedback than @dohadeer. Stats is fun.

Edited March 13, 2020 by mrniaboc

[Orraloon]

14 minutes ago, mrniaboc said:

Is it? Please do explain my major mis-step. I took a sample of the population who are regularly monitored for their health and also at a level of fame that means their ill health is reported, and showed that 1.3% of them is the fraction with coronavirus. 

I guess you could be arguing that professional footballers are more likely to contract the disease. But I can't see why at the moment?

It could be all that hugging and kissing that goes on nowadays. Filthy habit.

[Orraloon]

16 minutes ago, ParisInAKilt said:

Wonder if they are in the process of creating extra NHS 24 jobs, temporarily to deal with the virus. I know they are doing something similar in Australia. 

Yes, they started doing that a couple of weeks ago. 

[phart]

9 minutes ago, mrniaboc said:

Definitely can be considered a non-representitive sample for sure. But at the moment it's the only sample I have my hands on where I think all cases are being reported, and reported early.

Also defs small number stats that I'm grafting to a large population, so the numbers have large error margins.

But it was a fun thought experiment to do. Say 50% of footballers on london had covid. What estimate would you then make for the general London population? 

Thanks for giving better feedback than @dohadeer. Stats is fun.

None. I'm in the group that doesn't know how to sample.

[mrniaboc]

1 minute ago, phart said:

None. I'm in the group that doesn't know how to sample.

Hahaha. I'm in the group that likes to do back-of-the-envelope estimations that get some people very worked up.

[phart]

2 minutes ago, mrniaboc said:

Hahaha. I'm in the group that likes to do back-of-the-envelope estimations that get some people very worked up.

I used to get worked up a lot when younger my soul is much more serene these days.

[dohadeer]

1 hour ago, mrniaboc said:

Some quick statistical sampling:

Number of players + coaches of London EPL teams ~150

Of those there are 2 with confirmed covid 19

That's 1.3%

London's population = 9 million

Estimate of currently infected number in London therefore ~ 120,000

UK science chief reckons there could be up to 10,000 cases UK-wide. I reckon he could be wrong by at least an order of magnitude. 

Permission to use your post as a practice question for statistics students?

’Explain three errors that mrniaboc has made in their calculations.’

 

1. They are using someone else’s data - a mixture of media reports and estimates - rather than data they have gathered themselves.

2. They haven’t used a random sample of the population - they have chosen 150 people in the same industry, who have been in close contact with each other and had similar travel movements to each other over the past month.

3. They have started from a known result (2 people who work for Premier League football clubs in London, with the virus) and worked backwards, then used that to extrapolate the result to the whole population. That would be like knowing that Tom Hanks has the virus and then deciding to do an analysis of virus contraction amongst 1990s Best Male Oscar winners, and extrapolating that result to the whole population. You shouldn’t start from a known result and work backwards, a random sample would be a far better approach.